Upload
mite
View
221
Download
0
Embed Size (px)
Citation preview
8/6/2019 CS_m1
1/8
System Analysis and Design/Information for Management Case Study
V. Rajaraman/IISc M1/V1/July 2004/1
INFORMATION PROCESSING FOR A STORE - AN OVERVIEW A Small
Case Example
In what follows we will present a broad overview of how data is processed to meet the
functional requirements of a store. The presentation below will be an overview.
The important functions of a store are:
to keep an up to date ledger containing stock positions, cater to requisitions for issue of items from the store, initiate reorder of items whose stock is below a specified limit, update stock register when items are received, and answer enquiries regarding availability of items in stores.
In a computer based system the stock ledger is organized in a suitable form for easy
updating and retrieval and recorded on a magnetic disk. Magnetic disk storage is the
primary storage medium for storing large data bases. This is due to the fact that any
record can be directly accessed. Magnetic tapes are used primarily as a back-up storage
unit for keeping copies of data on disk. Tapes are also useful for storing old files and for
interchanging files between different computers.
In order to create a stock ledger for a computer-based system it is necessary to first assign
unique codes for each item in the store. The unique code assigned to each item is known
as the key of the item record and identifies the record. After that it is necessary to find
out what data fields are needed for each item in the stock. The fields are determined by
working backwards, that is, first asking what outputs are needed and based on that
8/6/2019 CS_m1
2/8
System Analysis and Design/Information for Management Case Study
V. Rajaraman/IISc M1/V1/July 2004/2
determining the data needed. The fields are organized as a record and stored in a data
base. This data base is the primary or master file for the store. (It is the computer
readable version of a stock ledger used in a store). Once the record format for each item
in the store is determined, one record is created for each item in the store. These records
are entered manually by a data entry operator who enters the records using a keyboard of
a terminal connected to the computer. This is called on-line data entry. Data may also
be entered on a separate computer such as Personal Computer (PC) and stored on a
floppy disk. This is called off-line data entry.
In off-line data entry the data entry machine is a low cost machine. If the volume of data
to be entered is very large, then a number of machines can be used and data prepared,
checked and corrected. As opposed to this, an on-line data entry method uses terminals
connected to the computer. In such a case the computer should be timeshared. On-line
data entry is appropriate for inserting, deleting or correcting some records in fields.
The data on floppy disk can then be transferred to the disk connected to the computer.
Off-line entry is used when the data base is very large and the computer used is a server
or a mainframe computer. If the data base is small, the PC itself may be used for data
entry and for data processing.
Before data is stored in the disk-file it must be ensured that any errors made during data
entry is detected and corrected. This is done by a program called an edit program and a
8/6/2019 CS_m1
3/8
System Analysis and Design/Information for Management Case Study
V. Rajaraman/IISc M1/V1/July 2004/3
control total checking program. Such programs are essential to ensure the validity of data
in a master file.
A procedure similar to the one used to create the master file is also used to keep the data
in the master file up-to-date when new items are received or new stock of items already
in the ledger are received. Table 1 summarizes the operations performed for other
functions. The format of a record for entering requests uses the same item codes assigned
in creating the master file. Other fields are determined based on what outputs are needed.
In this case a reasonable format for requests is:
(item code, item name, quantity requested)
Table 1 Operations Performed in Stores Information Processing
1. Create stock ledger
Codify items Determine data fields needed for each item Create a record for each item Organize the records as a data base
2. Issues/Reorder Codify items Determine data fields required in each request Determine data fields required for each issue Create record format for requests and issues Create record format for reorder
3. Receipts
Codify items Determine data fields required in each receipt Create record format for receipts
4. Enquiry Codify items Record format for enquiry Record format for response
The record format for a reorder request would be: item code, item name)
8/6/2019 CS_m1
4/8
System Analysis and Design/Information for Management Case Study
V. Rajaraman/IISc M1/V1/July 2004/4
( The quantity to be reordered is normally stored in the file maintained by the Purchase
department)
The record format for issues would be : (item code, item name, quantity requested,
quantity issued)
The request record is entered through a keyboard by a requester. It is checked for
validity by a program. A valid request is forwarded to a retrieval program which
retrieves the data on the requested item from the Master file on disk. The item code in
the request is used as the key for retrieval. If the requested number of items is available,
then an issue slip is printed; otherwise a reorder request is printed. Care must, however,
be taken to ensure that once a reorder is requested it is not requested again till the item
reordered is taken into stock. Each request slip processed by the computer is called a
transaction. If each request is processed as and when it arrives and the terminal on which
the request is entered is connected to the computer, then the processing method is called
On-Line Transaction Processing (OLTP).
There is another method of processing requests. A number of requests arriving during a
day (for example) are collected and formed into a batch. The data in such a batch can be
keyed-in off-line and a floppy disk created. This floppy disk can then be used to enter
requests on the computer which has the Master file. The entire batch is processed and
outputs are printed. This mode of processing is called batch processing. Batch
processing is usually more efficient. It, however, is not as timely as on-line processing.
In operations such as payroll processing which is done periodically, batch processing is
more appropriate.
8/6/2019 CS_m1
5/8
System Analysis and Design/Information for Management Case Study
V. Rajaraman/IISc M1/V1/July 2004/5
Enquiry system is normally an on-line transaction processing system as it is the most
natural way of answering users' queries. In an enquiry system also, a user's query is first
checked for validity of item code, etc. before it is processed.
A variety of information systems used in practice are primarily on-line transaction
processing systems. Common examples are airlines and railway ticket reservation
systems. Designing such systems require special care to ensure that response to enquiries
are fast and that the system has a hot standby if there is a failure. High reliability is
required as failures can be catastrophic (imagine many persons getting the same berth
reserved on a train). Similarly reliability and availability is essential in on-line banking
systems.
MIS and DSS for Stores
The processing methods presented in the last para are for routine data processing. The
information they provide is operational information. The system required to obtain
tactical information require further processing. Such systems are known asManagement
Information Systems (MIS). In the stores processing case study, some tactical decisions
would be: at what stock level should reorder be initiated? How much should be
reordered? These are determined based on data such as rate of issue of each item, time
needed for delivery from date of order, transport cost, storage cost, shelf life, and loss
incurred if an item is out-of-stock. These data have to be collected separately over a
period of time, often as a byproduct of a routine data processing system. In the stores case
daily issues of some critical items can be abstracted and the average issue can be
computed. Data on delivery times, transport cost etc. can be separately collected. Well
8/6/2019 CS_m1
6/8
System Analysis and Design/Information for Management Case Study
V. Rajaraman/IISc M1/V1/July 2004/6
known methods of operations research can then be used to compute stock level for
initiating reorder and the optimal quantity to be reordered.
Operational data collected over a period of time is called data archives and the process
of collecting it is called data archiving. With the availability of massive disks in which
terabytes (1012 bytes) of data can be stored, it has become feasible to analyse the
archived data. Analyzing archived data to observe patterns which assist in management
decision making is called data mining. A stores manager may, based on his experience,
think that in the months of October, December and April the sale of sugar is very high
compared to other months. This conjecture maybe verified by data mining. In data
mining a rule is formulated which may say that in October sugar sale is 1.5 times the
average, in December it is 1.3 times normal and in April it is 1.4 times normal. This rule
may be verified within a specified margin of error by examing the data archive. If the
rule turns out to be correct, a manager will be able to decide how much sugar is to be
stocked in these months. This is a simple example of the use of archival data and data
mining to assist in tactical management.
As another example of tactical information requirement, let us consider the question of
fixing credit limits for customers. In order to arrive at this, the following inputs would be
useful:
Customer details such as income, occupation etc.
Customer payment history
Volume of purchase by customer Outstanding dues (if any) from the customer.
8/6/2019 CS_m1
7/8
System Analysis and Design/Information for Management Case Study
V. Rajaraman/IISc M1/V1/July 2004/7
Using these one may formulate some rules to arrive at the credit limits and also predict
their possible effect such as:
anticipated effect on sales caused by varying credit limits anticipated loss/profit due to credit limits.By analyzing the impact of credit limits as specified above, a decision may be arrived at
to fix credit limits.
The primary point to note is that one has to formulate a model and sometimes simulate a
system to obtain tactical information. The operational information from routine
processing becomes an input to obtain tactical information.
Strategic information is obtained through what are known as Decision Support Systems
(DSS). In the stores example a strategic decision would be to reduce variety in inventory
by discontinuing some items in store, deciding what new items to introduce in the store,
and when to open a new branch. Decisions such as these require provisions for a variety
of data transformations and representations.
Strategic information is often unstructured. Strategic decisions are made after trying to
answer questions such as "What will be the profit if I take a decision and what will be the
long range loss if I don't take it?". In a complex decision many parameters will be
involved. Identifying these and predicting their impact on a decision needs judgements
coupled with analysis. For example, taking a strategic decision of whether to open a new
branch or not would require the following information:
Projected demands in the new branch Impact on current branch Pricing in new branch
8/6/2019 CS_m1
8/8
System Analysis and Design/Information for Management Case Study
V. Rajaraman/IISc M1/V1/July 2004/8
Competition in new location Profitability of new branchThese have to be provided using analytical and simulation model known as decision
support models. These models are more difficult to evolve than those needed in tactical
information development. Decision support systems should also provide aids to the
manager for conceptualization such as charts, graphs, etc. They should also provide
facilities to ask a variety of queries on the data base. A variety of summary reports
should be made available on request. The overall purpose of decision support systems
is to aid in strategic, unstructured decision making. Developing such systems is much
more difficult than developing operational systems. They, however, are the ones
required by the top management of organizations.